8 January 2020

Introduction

About me

  • MSc in GIS at UCL
  • PhD / Researcher at King’s College London
  • The London Hybrid Exposure Model / Air quality GIS ‘stuff’

  • Now at Guy Carpenter (Model development, Re-insurance)

Why measure air on the tube

  • Exposure to particles on subway systems > important

  • Seaton et al 2005, but …
    • Tox. mechanisms
    • Susceptible populations
    • Analytical techniques

Aims

What we tried to do

  • Measure variations in PM2.5 between lines and stations
  • Characterise the chemical composition
  • Calculate calibration factors for optical instruments
  • Provide a spatially resolved dataset for future analysis

Method

Mobile Measurement campaign

  • TSI AM510 SidePak (PM2.5)
  • Philips Aerasense (numbers and size of particles)
  • 31 hours
  • All lines
  • 89% of stations (NE Central, SW Piccadilly)

Geo-tagging data

  • Need to link air quality measurements to locations
  • No GPS signal on large sections of the network
  • Considered using timetables / interpolating between known locations
  • Ended up using a notepad

Characterisation & Calibration

  • Tricky installation at Hampstead
  • Particles collected on filters over 5 days measuring composition & amount
  • High time resolution equipment installed
    • Aethalometer / TSI Dustrak / 2 TSI Sidepaks / Micro-aethalometer

Passenger-weighted stations

  • 2015 tap in/tap out, Underground performance report
  • Annual in/out for each station
  • Mean PM2.5 measured at each station
  • Passenger rank * air quality rank = passenger-weighted ranking

Spatial representation of the tube

Results

Calibration factors

  • Linear model to calculate correction factors for mobile monitoring equipment
  • Mobile monitoring equipment co-located in tube station v. outdoor

The Victoria Line

Line averages

Station depth 1

Station depth 2

Depth on the Central Line

PM2.5 Map

PM2.5 online map

Passenger-weighted stations

Origin-Destination matrix

Characterisation

Conclusions

Conclusions

  • Particles tend to be larger in diameter than those at background or roadside environments
  • More particles
  • PM2.5 varied between lines & locations
    • lowest Hammersmith & City (Mean 25 µg/m3), similar to roadside
    • highest Victoria (381 µg/m3), 15 x higher than roadside

Conclusions 2

  • General relationship between ‘depth’ and air quality
  • Oxford Circus, Waterloo, London Bridge, Victoria and Vauxhall at top of passenger-exposure ranking
  • 79% of PM2.5 characterised
    • 47% iron oxide, 7% elemental carbon, 11% organic carbon, 14% metallic and mineral oxides
  • Previous studies using light-scattering may under-report PM

What next

What was planned

  • Characterise the remaining 11%
  • More measurements accross the network to improve understanding
    • train frequency
    • passenger numbers
    • time of year
  • Interventions?
  • Develop inclusion in exposure modelling

Test

What happened

  • Map or two from GC work

The end

Publication, Contact & Data